J Manipulative Physiol Ther 1999 (Jul); 22 (6): 376–381 ~ FULL TEXT
Lynton G. F. Giles, DC, PhD, Reinhold Muller, PhD
National Unit for Multidisciplinary Studies of Spinal Pain,
Townsville General Hospital,
Patients referred to Townsville General Hospital outpatient Spinal Pain Unit in Australia for evaluation and treatment of chronic (>13 weeks) spinal pain were randomized to acupuncture, medication or spinal manipulation. After 30 days of treatment only the manipulation subgroup showed significant reduction in pain intensity. The manipulation group displayed uniform, significant, substantial improvements across all outcome measurements while in the two other intervention groups no significant improvement could be found.
OBJECTIVE: To compare needle acupuncture, medication (tenoxicam with ranitidine), and spinal manipulation for managing chronic (>13 weeks duration) spinal pain syndromes.
DESIGN: Prospective, randomized, independently assessed preintervention and postintervention clinical pilot trial.
SETTING: Specialized spinal pain syndrome out-patient unit at Townsville General Hospital, Queensland, Australia.
SUBJECTS: Seventy-seven patients (without contraindication to manipulation or medication) were recruited.
INTERVENTIONS: One of three separate, clearly defined intervention protocols: needle acupuncture, nonsteroidal anti-inflammatory medication, or chiropractic spinal manipulation.
MAIN OUTCOME MEASURES: Main outcome measures were changes (4 weeks vs. initial visit) in the scores of the (1) Oswestry Back Pain Disability Index, (2) Neck Disability Index, and (3) three visual analogue scales of local pain intensity.
RESULTS: Randomization was successful. After a median intervention period of 30 days, spinal manipulation was the only intervention that achieved statistically significant improvements (all expressed as percentages of the original scores) with (1) a reduction of 30.7% on the Oswestry scale, (2) an improvement of 25% on the neck disability index, and (3) reductions on the visual analogue scale of 50% for low back pain, 46% for upper back pain, and 33% for neck pain (all P<.001). Neither of the other interventions showed any significant improvement on any of the outcome measures.
CONCLUSIONS: The consistency of the results provides, in spite of several discussed shortcomings of this pilot study, evidence that in patients with chronic spinal pain syndromes spinal manipulation, if not contraindicated, results in greater improvement than acupuncture and medicine.
From the Full-Text Article:
The high incidence of chronic spinal pain syndromes, their recurrent nature in many patients, and their contribution as a main cause of absence from work is well documented.  It is a disconcerting situation that the undisputed immense public health impact of chronic spinal pain syndromes and their enormous cost for health care budgets world wide stand in sharp contrast to the sparse knowledge about absolute and relative efficacy of different interventions. Very little research with respect to thoroughly planned prospective studies has been performed to date. There is controversy in the literature regarding the most effective treatment, although Meade et al [1, 2] showed that when chiropractors and hospital therapists treat patients with low back pain as they would in day-to-day practice, those treated by chiropractic practice derive more benefit and long-term satisfaction. The trial by Meade et al,  proclaimed to be “one of the better trials in this field,”  was a “pragmatic” trial that tested what happens in day-to-day practice and in which details of the type, frequency, and duration of intervention were at the discretion of the treating clinician. However, a “pragmatic” trial may not identify the components of the more successful intervention that were responsible for improvement, but a “fastidious” trial may. 
To shed some light on the comparative efficacy of common treatment regimens, a prospective, randomized, independently assessed preintervention and postintervention clinical pilot trial was established at the specialized spinal pain syndrome outpatient unit at Townsville General Hospital. In this “fastidious” trial approach, three separate and clearly defined intervention protocols,
(1) needle acupuncture alone or in conjunction with low voltage electrical stimulation
(depending on patient tolerance),
(2) spinal manipulation, and
(3) nonsteroidal anti-inflammatory medication (using tenoxicam with ranitidine)
were tested to compare the efficacy of these standardized intervention protocols.
At first glance the high proportion of dropouts, which additionally differed significantly between the intervention groups, seems to be a major drawback of this pilot study. It must be emphasized, however, that these dropouts were not related to major improvement or deterioration or side effects (only 3 had side effects) but were simply caused by severe understaffing and the resulting restrictions in the follow-up in an environment where major traveling was included (the catchment area for patients of the Townsville Spinal Pain Unit reaches as far as 1000 kilometers inland). All dropouts were telephoned at the end of the study to ascertain why they did not return. Because all reasons were independent of the outcome of the intervention, the dropouts have no potential for confounding the results. The presented results are also not confounded by any of the specifically recorded potential confounders (eg, sex, age, pain duration, and number and type of previous interventions) because bivariate findings were checked for confounding and found to be invariable.
To the best of our knowledge, these are the first findings reported from such a randomized clinical trial, and they seem to be consistent with the findings of the “pragmatic” study of Meade et al ,
 that spinal manipulation has an important role to play in the treatment of spinal pain syndromes. The complexity of the field of spinal pain syndromes, as well as political obstacles associated with a multidisciplinary trial, are impediments in this field of research, but they can be overcome, as this pilot study shows.
The most remarkable feature of the results of this pilot study, however, is the absolute consistency: over all outcome measures, the manipulation group displayed the most substantial improvements that were uniformly found to be significant. In the two other intervention groups, not a single significant improvement could be found in any of the outcome measures. The only other significant results were found in the whole study group. However, a close look at the results easily clarifies that the significant changes in the whole study group are simply a weaker reflection of the more substantial improvements observed in the manipulation subgroup.
It should be mentioned that the relatively short-term nature of this pilot study cannot indicate a guaranteed long-term benefit. The fact that patients could be discharged does not necessarily imply permanent relief from symptoms. To assess long-term effects, additional data collection over several years would have been necessary but was beyond the scope of this pilot study. Nevertheless, it is a noteworthy finding that significantly more patients could be discharged after spinal manipulation as opposed to the other interventions. This finding is consistent with and therefore corroborates the results found in the subjective outcome measurements.
Because a pathologic cause cannot be identified for most episodes of spinal pain,  the financial and resource implications of low back pain alone are extensive.
 Therefore clinical guidelines are needed to improve the management of chronic spinal pain in primary care
 because the proportion of patients with spinal pain who have poor outcomes seems to be higher than generally recognized
; research into this major health care problem is essential. At a time when the whole clinical community is implementing evidence-based management of patients, it is imperative that the field of chronic spinal pain syndromes, which vastly impacts on patient suffering and scarce health care resources, should catch up with other health care programs and no longer be characterized by a disproportionate paucity of epidemiologic data and controversy in comparison to its overall impact. In spite of the documented and discussed shortcomings of this pilot study, the consistency of the findings has enabled some light to be shed on the benefit of different interventions on this costly and debilitating condition.
Major well-funded prospective longitudinal clinical trials should be undertaken as soon as possible to further investigate this complex issue. These studies should emphasize specifically (1) standardized intervention protocols such as those used in this study, (2) strict monitoring and follow-up of patients in each treatment category, and (3) the long-term effects of treatment that should be evaluated by means of 12-month questionnaires administered over at least 3 years, which was beyond the scope of this pilot study.